Title of Invention

"SOFT DECISION ENHANCEMENT"

Abstract A Vierbi trellis processing technique in which soft decisions and hard decisions arc derived from a received signal and the soft decisions are enhanced by being modified using the hard decisions. A log likelihood ratio for a bit of the received signal can be derived by grouping candidate metrics associated with the decision that the bit has a first state, grouping candidate metrics associated with the decision that the bit has a second state, applying respective functions to the groups and calculating the difference of the function values.
Full Text i
!.. mobile communication systems, u transmission link between a transmitter and J rec-'-ei .'¦¦:¦ ..-uiM-r fro»- a number of impair;, ic;its. Two such effects are thermai noise and niultipath
I-.]i:liipath fading um result in Intersymbol Interference (ISI) at the receiver if the delay spread uf ;L:O channel is larger than the modulation period ('Digital Communications', Proakis, 2nd IviJuon. McGraw-Hill). Hence, in a given propagation environment, ISI will become more of a yiobLm as uu !r:i:ismission rate increases. For communication systems aimed at providing medium to high data rate services, the presence of ISI can severely limit the throughput of a tr?;r.:mission link and degrade the quality of service experienced by a user.
Scne digital communication systems also introduce ISI by design. This is the case in the E-GPJIS system., where the modulation pulse shape used to improve the spectral efficiency of the in ¦ -rutted signal generates ISI. For more information, see 3GPP TS 05.04 V8.4.0 (2001-11), A' !'hiV;il Specification 3rd Generation Partnership Project; Technical Specification Group vJ'M/LvDGE Radio Access Network; Digital cellular telecommunications system (Phase 2+); \\ ii.'j!aiion.
. -]'¦• "jr. i .-inuriY, uf performance degradation experienced by a user of a Time-Division Multiple
\..\.\;ss (TDMA) cellular communication system is interference generated by other users in the
_, .j^'ict using the same carrier or adjacent earners. These interference effects, referred to as co-
chsi.nel interference and adjacent channel interference respectively, can greatly reduce the
Nc: \:ity of a cellular system.
. .'¦: .,{' L;2 imp: .r, :JIUS described above make it difficult for a receiver to reliably recover
„. ;>rmntion thai :i transmitted signal iu crded to convey, leading to the use in receivers of
..--Miplex algorithm^ for demodulating received signals. The implementation compicxiiy of
. ."iy..rithms \vi;i have a significai\ inijiact on a digital receiver in terms of overall silicon

2
lie size, process ¦:>> speed, power consumption and memory requirements. Hence, the use of an iTicient receiver architecture offering good transmission link performance is of considerable
.uportance.
\<: order to improve reliability of a communication link forward error correction> coding can be embedded in a transmitted signal. An FEC coding operation introduces
redundancy to n transmitted signal and this redundancy can then be used at a receiver to
improve the accuracy of estimates of the transmitted data that are generated by the receiver.
i-fowever, for FEC coding in a transmitted signal to be most beneficial, it is important that such
a signal is demodulated by a receiver in a format which can be best interpreted by an FEC
decoding process within the receiver. A number of such receivers have been proposed in the
;:ast. See, for example, 'Optimal decoding of linear codes for minimizing symbol error rate',
. Bahl, J. Cockc, F. Jelinek, J. Raviv, IEEE Trans, on Information Theory, Volume: 20,
larch 1974; 'A Viterbi algorithm with soft-decision outputs and its applications', J.
agenauer, P. Roeher, GLOBECOM'S9, Dallas, November 1989; 'On the equivalence
;;tween SOVA and Max-Log MAP decodings', M.P.C. Fossorier, F. Burkert, S. Lin and J.
-Jagenauer, IEEE Communications Letters, vol. 2, no. 5, May 1998; 'Soft information in
oncatenated codes', B. Rislow, T. Maseng, O. Trandem, IEEE Trans, on Communications,
ol. 44, Issue 3, March 1996; and TCM on frequency-selective fading channels: a comparison
T soft-output : oabilistic equalizers', P. Hoeher, GLOBECOM '90., December 1990.
nvever, the im;. lamentation complexity of such prior receiver architectures is usually Ii-i_rh.
:: order to describ. the computations performed by-r receiver, it is first necessary to present a ,.:>del of the transmission link in which the receiver participates. A suitable model is
.esented in Figure 1 and will now be described in some detail. This model assumes that the iormation is transmitted in blocks of bits rather than as a continuous stream. However, it
hould be noted that the invention described later in this document is applicable to both types v f transmission.

3
According to the model, a transmission block {w*}^/, K is made of A'"information bits, ?v. e {0,1}. Error protection is added to these bits in order to improve the reliability df'ffte
transmission. The error protection coding unit 101 generates from the block {uk)>ich ,., a
transmission block {d,.}^ D) made of D (where D > K) information bits,^. e{O,lj. The
information bits d>c are grouped into C sets of M bits (it can be assumed, without loss of generality, that D - M x C). Each of these sets can be denoted A/o where
Each set of PI information bits Ak is modulated onto the complex plane using a modulation scheme M that maps sets of M bits onto the complex plane. This mapping is performed by a modulation unit 102. For example, in the case of an SPSK modulation, the modulation M can be expressed as:

A slightly modified version of the 8PSK modulation described by the above equation is used in uie E-GPRS system.
A set of M information bits \dyxk,...,d^Mxk^M_A can be uniquely identified with a single
decimal number / l0

4
This equation maps the M binary values in a set A^ to a unique value in the set|0,...,2w~'j The set of information bits db [b e {0,...,M}) that verify the above equation for a given value
-. 7 can be denoted £>:' (/) (b e {0,...,M}).
¦ ;ie C modulated symbols representing a transmission block are transmitted over the air and in
;,a process are distorted by the propagation channel 103. Assuming a general model with
:n;mory for the propagation channel, the samples {sk} . , at the input of a receiver can be
. .pressed as: .k=F(c4,&_,)
' I:re, ck =M(Ai) and £k represents the state (memory) of the propagation channel when the
¦ modulated symbol is transmitted. Note that any filtering performed by either the transmitter and/or the receiver can be incorporated in the propagation channel model. The
Lappings F and S used to model the propagation channel can be time varying. However, to 1 implify the notations, it is assumed in this document that these mappings do not depend on time. Note, however, that the approach described here is also applicable to time-varying i^annels.
In most cases, the propagation channel mappings F and S can be modelled as linear filtering operations:


In the above example, it is assumed that the memory of the channel is limited to L modulated symbols. In reality, the memory of the channel may be infinite. However, for any desired power threshold T, it is usually possible to find a value L such that:


|/?,f )£-!

Hence, by selecting the threshold Tsuch that the residual power is low enough, it is possible to
--.ssume that the channel memory is limited. When this is done, the channel mapping can be
ascribed with just a set of filter coefficients {ht} ..
•- the model described here, it has been assumed, without loss of generality, that the filter representing the channel propagation is causal.
la order to recover the transmitted symbols Ck, the receiver will need to know the propagation
^annel mapping. However, the receiver will usually not have prior knowledge of the
propagation channel conditions. It may nevertheless be possible for the icceivei (o generate
estimates of the channel coefficients \h,\ , . which can be used in place of the true values.
For example, in the EGPRS system, a transmitted signal will incorporate a pattern, referred to ;J:; training sequence, which is known to the receiver and the receiver can use this training sequence to generate an estimate of the propagation channel conditions.
The signal is,}, , , is corrupted by additive in.) noise, introduced at process 104,
such that a block of symbols from which the receiver tries to recover transmitted information can be expressed as:
rK:=sL+nk (ke{l...,C))

The aim of a receiver is to generate the sequence \ut \k which maximises the set of A-
Posterion probabilities P(ui\R), where /: belongs to tne set{l, . . . ,C) and R denotes the
sclKL{1....,cr
Note that if one assumes that the information bits take the values 0 and 1 with equal probability, then the maximum A-Posteriori solution is also the solution which achieves the maximum likelihood criterion:
TDSX.P(R\UL)
Hence, a full maximum A-Posteriori receiver will estimate a sequence of transmitted bits ("*}/-e{i K) ^ taking into account the error correction coding and the modulation scheme at
the same time. Such a receiver will perform well and allow a good throughput for a transmission link. However, because error correction decoding and demodulation are performed simultaneously, the implementation complexity of such a receiver is almost always prohibitive. Hence sub-optimum solutions are usually preferred. Practical solutions usually separate the demodulation from the error correction decoding, as in Figure 2. The sequence of
received symbols \r,.\. . „, is used by unit 201 to calculate estimates \d..\ of the
transmitted coded bits {^t}te,, D)- The error correction decoding unit 202 uses the symbols
\dk in order to derive estimates iu,.), . „",' of the transmitted uncoded bitsjwj.1, „ ,,.
'. M (.e(i,. ,,D) l 'H* K) ( klHi--*)
The algorithms used for the error correction decoding 202 can be divided into two broad classes depending on the type of inputs being used. The first class corresponds to receivers which use hard decision decoding. These decoding techniques use only the estimates of the
coded bits Id A in order to produce estimates fw,.}, ., „. of the uncoded bits. The
recond class corresponds to receivers which use soft decision decoding. These decoding

7
technique rely not only on the sequence cf estimated bits UA but also make •
( )lce{l...,D)
information about the reliability of this bit sequence. Because the soft decision dec. ¦¦.';•
approach uses information about the reliability of the estimated bits d, \ , soft decisic
I )ke(l £>)
decoding receivers usually perform better than hard decision decoding receivers.
When the soft decision decoding approach is used, the information input to the error correction decoding unit 202 is the Log-Likelihood Ratio (LLR), which for the coded bit dk is expressed as:
'p(dk=tl\R)'
As a side-note, it can be seen that estimates of the coded bits can be generated from these LI ' using the following rule:
• =(0 if Ak0
Note that in the case when/lj.=0, both dk~0 and dk—\ are equally likely, hence c;.
decision can be made.
According to one aspect, the invention provides apparatus for manipulating a received si^n.' which corresponds to a transmitted signal comprising a series of bits, the apparatus comprising means for processing the received signal using a Viterbi algorithm to produce ca;vJ: ¦'>'¦¦ metrics for states arranged in a trellis and to make selections from amongst the canc'iV! metrics to provide metrics for said states, means for producing an initial soft decision fo: a b, of said series from said candidate metrics, means for producing a hard decision for said :
initial soft decision in a manner dependent upon, the hard decision to produce an enhance..1" indecision for said bit.
According to that aspect, invention also consists in a method of manipulating a received signal which corresponds to a transmitted signal comprising a scries of bits, the method comprising a step of processing the received signal using a Viterbi algorithm to produce candidate metrics for states arranged in a trellis and to make selections from amongst the candidate metrics to provide metrics for said states, a step of producing an initial soft decision for a bit of said series from said candidate metrics, a step of producing a hard decision for said bit from a metric selection corresponding to reception of said bit and a step of modifying the initial soft decision in a manner dependent upon the hard decision to produce an enhanced soft decision for said bit.
In certain embodiments, the initial soft decision is produced as a log likelihood ratio from candidate metrics.
In certain embodiments, the hard decision is produced by storing a decision history of metric-selections that are made in calculating metrics through the trellis and then tracing back through the decision history to produce the hard decision. The trace-back operation may be mi li a too after processing only a portion of the trellis.
ii • ortaih eruLodiments, metrics are produced for the trellis states using a reduced-.-; iu technique in which metric selections made in moving through the trellis are treated as symboi decisions for calculating metrics further along the trellis and one of the metric selection^ is selected to provide the hard decision. The metric selection may indicate a modulation syml >; that requires translation to recover the hard decision.
IJI certain embodiments, the value of the enhanced soft decision is selected from a group values conditional on the state of the hard decision.

In certain embodiments, the value of the enhanced soft decision is selected from a grout values conditional on the polarity of the initial soft decision.
According to another aspect, the invention provides apparatus for manipulating a recc:... signal using a Viterbi algorithm to produce metrics for states in a trellis, wherein the recc:. ;.vi signal corresponds to a transmitted signal comprising a series of bits and the apparatu. comprises means for performing an update of the metrics corresponding to reception of a h by calculating candidate metrics for all the states, means for calculating a function of a plurality of the candidate metrics that correspond to a decision that the bit is a logical one. means for calculating a function of a plurality of the candidate metrics that correspond to a decision that the bit is a logical zero and means for calculating the difference of the values of the two functions.
According to that aspect, the invention also consists in a method of manipulating a received signal using a Viterbi algorithm to produce metrics for states in a trellis, wherein the received signal corresponds to a transmitted signal comprising a series of bits and the method comprises a step of performing an update of the metrics corresponding to reception of a bit by calculating candidate metrics for all the states, a step of calculating a function of a plurality of the candidate metrics that correspond to a decision that the bit is a logical one, a step o.' calculating a function of a plurality of the candidate metrics that correspond to a decision ihr. the bit is a logical zero and a step of calculating the difference of the values of the r.w functions.
Received bits used in progressively calculating metrics through the trellis may arrive as pan ot a modulated symbol conveying more than one bit of the transmitted signal. It may be the ca^-that modulated symbols rather than received bits per se are used to update trellis metrics.
The routines according to the present invention for manipulating metrics and soft decision: can be implemented in hardware or as software for execution by suitable data pnxx\;

10
hardware. These routine:; can implemented within, for example, a base station of a v/; communications network or a telephone for use in such a network.
The manipulation of the received signal by the invention may be for the purpc-demodulation or decoding of the signal. In the former case, the trellis states may by defin,.; the possible permutations of symbols in the channel memory. In the latter case, the tr: states may be defined by the possible permutations of symbols in a memory arrange;: forming part of a convolutional encoder involved in the production of the received signal.
By way of example only, certain embodiments of the invention will now be described with reference to the accompanying drawings, in which:
Figure 1 illustrates a model of a communications link; Figure 2 illustrates a genre of receiver architectures;
Figure 3 illustrates a full-trellis demodulator scheme according to an embodiment of the invention;
Figure 4 illustrates a reduced-state trellis demodulator scheme according to an embodiment the invention; and
Figure 5 illustrates a receiver architecture suitable fci hosting a demodulator scheme accoaii: to the invention.
The proposed receiver architecture also has the general structure shown in Figure 2. i demodulation unit 201 uses Viterbi processing to produce a stream of soft decisions, in ihc form of LLRs or estimates of the LLRs, for use by the error correction decoding unit 202. T different computations involved in the Viterbi processing will not be detailed and only tho.>e computations that facilitate the description of this embodiment of the invention will now he

presented. For a more detailed description of the computations involved in the Vitrv:--' processing used here, see, tor example, TCM on frequency-selective fading channel::-: comparison of soft-output probabilistic equalisers', P. Hoeher, GLOBECOM '90, December 1990.
The implementation of the Viterbi processing used in the demodulation unit 201 will first b-described for the case of a full trellis (i.e. all the hypotheses for the transmitted bits associatt a with (he different channel taps are taken into account); the case of a reduced-state trellis v. be described later. In the full trellis case, the number of states in the trellis used to perform tlu Viterbi decoding, denoted S, is equal to 2MxL. Moreover, any given state in the trellis has 2' "merging states" leading into it.
When the symbol rk is received, the metrics associated with the different states in the trail': are updated according to the equation:
rm(/c + l)=--;=maxJ^+1(0} 0 Note that, for a given modulation M, there is a one-to-one relationship, denoted H(»z), between each state index m and the set of hypotheses {ck,...,ck_L+2} for the transmitted information symbols. H is a function that maps a value from the set (0,..., S -1} to a value in a set of L-l complex numbers.
Also note that it is possible to add or subtract any value, constant over the different states, to this update without changing the soft decisions generated by the algorithm. Such an approach, may be used when the proposed receiver architecture is implemented using fixed-point numbei representation in order to reduce the computational complexity.

The candidate metrics %*+s (?) are calculated using xhe state metrics from the previous iteration with ths recursive equation:
^hl{0 = ^»,)W--B(/i.p(»».').fl')
Here, p(m,i) |0 The branch metrics are calculated using the following equation:
ti(rK,p(m,i),m) =
u=0
As a full trellis approach is being used, the hypothesis samples ck_u used in the above br:1; metric calculation are taken from the statesp{n:,i) and m that the bratich metric interconnects
>-'>~U: tbv.i the branch metric described in the above equation can be calculated in other WL/r;. • pc-^'s i:.- :¦- modify the computation of the branch metric, based on, for example, knowioii oi. or an 21: '¦,,. raion of, the statistical distribution of noise affecting ihe transmitted block
The LLR for the set of transmitted bits jd, v —,d/lfix,k_uiuu_l are then calculator', the following equation:
ie{o,....2l/-l|c;'(/)«[} iejo 2J'-lJD;'(/)=o}

13
The various computations presented above .n relation to Yiterbi processing coiTespon ' The foregoing equation for calculating the LLR AMx,k_L+]\+b only uses the maximum
metrics z':J' (/) corresponding to the decisions Db~l (i)~0 and D'b[ (/')=1 for the bit clt
decoded. The LLR can be improved by grouping first all the metrics corresponding decision 1, then all the metrics corresponding to a decision 0 and calculating the differ This is shown in the following modified equation:
cai:_.

hi
^-unW (*:+I(0)- / (*:+'(0) &e{0,...,M-l}
me{0, ...S-1} »ie{0 S-l}
(e{0 2A'-l|D6-'(,)=l} ,e{o,..,2"-l|D;'(i)=o}

/is a function which takes as input S x 2M'[ candidate metrics and returns a single value \ .>,. that a non-linear processing step can be used for the combination, using function/ of fk; metrics within each group rather than a simple a summation. One possible implementation of the metrics combination function/is as follows:
/(*c,x,,...,xSx.,,w_1) = log Z e 1=0
where a is a constant which can be derived, from example, from the estimation of the power of the noise in the received signal. Note that the computation of the above metrics combination function can be efficiently implemented by recursively using the following equation;

log [ea + e:>) = max (a, b) + log (l + e^*-*)-™- A ¦ j
The decisions made in selecting which caruiuiau metrics become the state metrics accc;- ¦-•: the equation ;-'ai(/c + l)= max U,'*+'(Oj lv/tiere 0 "initial" LLRs, generated using either of the two schemes described earlier, into "erJr
LLRs.
The decisions made in the calculation of the slate metrics needed to produce the initial Li Rs are stored in a decision history. For a given state metric ym (Jc +1), the information siore the decision history is the modulated symbol / to which the decision corresponds, as gr ¦ i :
the following equation:
rm(^ + l) = argmax{^+1(;)}
;=0 2"-l
Once a decision history has been accumulated for an entire received block {>'t}keil c], it is
possible to trace back through the trellis representing that block with the aid of the decision history to derive the path which corresponds to the most likely sequence of transmitted modulated symbols. Each trace-back step will generate an estimate a modulated symbol Ck which, in turn, generates an estimate of the set of transmitted bits
(^t-Lil.^wnim-'^w^-i) • Note that the 'mitial LLRs "e generated in increasing order of k in a forward computation phase, whereas, during the trace-back processing. !.e estimated transmitted bits {dMx{k_MydUxik_LA)i,,...Ju If the trellis is terminated to a known state, the trace-back processing can be performed starting from this known state. On the other hand, if the trellis can end with equal probabilities i:. r /¦'

15
of the states, then the trace-back can be performed from the state which the state metrics indicate as being the most likely.
Mote as well that it is not necessary to wait for the completion of the trellis processing of . Uitire burst before the trace-back computations a:e performed. The trccc-back processing car. oo performed at any point during the aforementioned forward computation stage. lr.. selection of the point at which trace-back processing is performed allows r..c. , -. requirements to be balanced against computational complexity.
'rV.-j imt.i.i L,iJ\. ^/,_/V*A computed during the forward phase are combined wul: •' . estimated transmitted bits dhutk_M\+b derived from the trace-back processing lo g'.ncnu enhanced LLR which are denoted X. ,. , .> ,.
V-\ combining the initial LLR values with the bit decisions dM,k_ulyb generated by the irac make use of the redundant information embedded in the received signal that the initii value do not employ. Moreover, the increase in complexity and memory requirements cause a by the trace-back and LLR modification processes is very limited.
Different approaches can be selected for the combination of the initial LLRs with i transmitted bits estimated during the trace-back process. Two possible approaches will no described.
In one approach, the enhanced LLRs are generated using the following rule:


I
XM*{k-u\yb

M«(k~L+\)+b

ifd =0
l/k(i-l+l)U

p
When this combination method is used, UT;- sign of the updated LLR values ahva-.s co-.rtspoixd*, .0 the MLSE solution. Hence, .'his proposed ]n another approach, the enhanced LLRs are generated using the following rule:
fo if (A,,,. , n .
'A/\(fc-£t-l)+6

0 if Uu /, , IUA>0andd =1
4/x(i-i+i)+i otherv/ise

Note that in the two methods of producing the enhanced LLRs described above, the association rule between the estimated bit dM,k_L+,^b and the sign of the enhanced LLRs can be changed to match the decision convention used by the receiver.
Figure 3 presents the different processing stages that are involved in the generation of enhanced LLRs from initial LLRs. The state metrics update unit 301 calculates the state
metrics ym (& + l) using the maxima of the candidate metrics of the braches that lead into the
ikutes. As part of these state metrics update computations, the various candidate metrics Xk+X {*) and the different best candidate metric indices Fm(& + l) are generated as described
earlier. The candidate metrics %M (0 are use^ by unit 302 in order to derive the initial LLRs \i*.ik-L+\s+b ¦ Once calculated, the initial LLRs are stored in an LLR memory 304. Similarly.
the branch metric indices Vm (k + \) are stored in a path history decision memory 303. T'v when the trace-back point is reached (for example at the end of the transmission bloclg. the

17
initial LLRs are, using one of the rules described above, combined in unit 305 with the decisions dKl^k_ux)+h derived from die trace-back processing in order to derive the enhanced
LLRs ^A.K(/.-L+I)+/> • Although it is possible to implement units 301 to 305 as separate pieces of
hardware, it is more likely that units 301. 302 and 305 will be implemented as software elements for execution by a microprocessor with the functions of memories 303 and 304 being served by a single memory associated with the microprocessor.
,\.s mentioned earlier, the demodulation stage jfthe proposed receiver architecture can employ reduced-state, rather than full-trellis, Viterbi processing. For long propagation channel.; u c. 1. lieie L is '-.rgc) or for communications systems using high-order modulation schei ..... .
where. M is. large), the implementation complexity of the full trellis approach is often prohibitive. For example, in the E-GPRS syslom using 8-PSK modulation, the total number OL s.ute.-;. ¦, -c,i or channels with relatively low delay spreads, is far too large to be elTi ion iy implemented in a mobile receiver. Different approaches have been proposed in the nnst in ¦'¦ d:-r ' .• -'.JU ihe number of states to be piocessed in the receiver. These appv. . generaiiy lei-med to a reduced-state processing, include, for example, the Delayed DeeiJion Feedback Sequence Equaliser (DDFSE; sec, for example, 'Delayed decisicn-il-. iback sequence estimation'. A. Duel-Hallen, C. Heegard, IEEE Trans, on Communications, Vol. Y!. Issue 5, May 1989) and the Reduced Stats Sequence Estimator (RSSE; see, for example. ' ¦¦educ'.;!-s:ere sequenc-j estimation for ceded modulation of iniersymbol interf. cnannels', ivl.V. Eyuboglu, S.U.H. Qureshi, IEEE Journals on' Selected Arc Communications, Volume: 7 . Issue: 6 , Aug. 1989).
A DDFSE architecture will now be described in which bit decisions dM,k_ux,Jrb are used to
produce enhanced LLRs. It will, however, be understood that the principles of the sell.':;: ^ described in this document for modifying LLRs using hard decisions about transmitted bits also be applied to receivers using different reduced-state Viterbi processing architectures as the RSSE.

Id
In a DDFSE receiver architecture, the state reduction is achieved by splitting the channel into two sections. The first section of the channel (with length L,) is processed in a way similar to
the full-state approach. However, for ihe remainder of the channel (with length Lr such that L = Lf + Lr), the modulated symbols derived from previous decisions are used rather than testing all the possible hypotheses. Using this method, the number of states for which trellis
processing is performed is reduced from 2iWx"~1^ to 2 ^ r '. To reflect this change in the
number of states, the branch metric computations need to be modified as follows:

B(rk,p(m,i),m) =

lr\ l; + Lr-rt-'2LKxck_u- 2 Kx^-u{p(mA)
¦1
H=0 u=L

M can b' sect: from the above equation that the computations involving the first Lf taps au {^juiicai '" \ iSv poi-jj^nr.cd in the fall trellis approach. Howcvci, lay the last 1. taos. the hypothesis symbols ck_v ai"e replaced by hard decision symbols c,._u (p(/«,/))• Thes^. ... -
decisions symbols are generated using the; decisions taken during the candidate metrics' si:'ectior. Tk; hard decision symbols are associated with each slate in the trellis ar.d sie updated c,s uie siaLe metrics are updated si;-'- that the selection involved in determining yr (k + \) detennines ck (m). Each hard decision symbol ck_u (m) is uniquely associated \vi '¦;


A/y(t-w4-l)+J)
/•/ bit decisions cl

b e {0,...,M-l}. The branch metrics are then used in one

scheme:; outlined earlier in order to generate initial LLRs ^Wxo._L+1x ,, ¦
The }'n(k + l) are compared for all m to deduce which state is most likely to be correct. Inherent in that state, is a set of hard decision symbols ck_H(m) whe u e \L j,..., Lj + Lr - 1/. The ha:d decision symbol ck_L_L+l(m) in that ¦
corresponds to M bit decisions dM,k_L+xy_b jo e |0,...,M-l}. Those bit decisions can then be

-onvouica Vv;tlithe njrda. LLRs \.
^,!+I (0 in order t0 derive the enhan^.1 ." ;' ',, ,(i_^;j+A using, for example, either ot ¦:. c
rules described above since the methods used in the full-trellis approach for the combination of the bit decisions with the initial LLR values can be used just the same in the DDFSE
architecture.
Note that, in the preceding paragraph, the earliest member ck_L _L+I(m) of the set ck_u(m) where u e \Lf,...,Lf + Lr-\) is used to create the enhanced LLRs AM,k_U].b. The set of hard decisions ck_u(m) where u e \L f ,..., Lf + Lr - l} is the smallest set of hard decisions
that need to be stored in order to perform the branch metric computations. Hence, by using hard decisions from this set in order to generate the enhanced LLRs, the memory requirements for the receiver are kept to a minimum. However, by increasing the amount of information stored in the receiver, it is possible to combine the initial LLR estimates v/ith older hard decision symbols. By so doing, the accuracy of the enhanced LLR values can be imprc ::± However, this introduces a larger processing delay in the generation of the enhanced LLRs.
llie i.iaiu t .L-LCJWICC ixu'een the processLn; performed for full-skit; and reduced si? ~ architectures lies in the way the hard 'i:c:::ions are generated. In a full-trellis receiver archi'ecture, all the branch metric decision; ai3 stored such that a trace-back operation c: \ be
p;u,.. ,i- ncc iho eonvplele transmission i.lock has been processed in order to derive '\, . ..A
i':cir ¦¦v- i'i ¦'. : ¦'ir.-.cd-statc receiver architecture, the hard decisions become available as the
:¦:; ¦ :, ;;i'e generated and therefore i;;c enhanced LLRs can be produced wh" ¦ .'1
J.L.r ; '!:>J:) is on-yoing. This means thai no explicit trace-back operation is required bra reduced-state receiver architecture. Note, however, that the set of hard symbc? ^cisions ck_,,[:n) .;.-: ocuitcd with the most probriNc ym(k + l) is analogous to a partial tra<:- k. moreover there is no need for a large path decision memory only the last li :v.rd symbols ck_u to be stored each state.>
20
It is. however, pcrsibls lo use a receiver archiiecture where a sequence of decision symbols extending hack beyond ihe last A, ;!'":•.'.•,,¦.>!¦ .>• :nbuis are stored for each state. This mean? that the partial trace-back represen'.ec by such 3 sequence will extend back earlier in time -i is well known, the further back a trace-back operation runs, the moie reliable its results wiil be. Therefore, the hard decision symbols towards the "early" end of such a sequence will be -tore reliable ihj arthtr bud in time the sec;; cix<: extends. hence b extending tlie lengih of the sequences hard symbol decisions an opportunity is created for using more reliable :r uibol lc at :cn enhance- however lengthening seqn.-.. attached to .states leads increa in amount memory required and delay that elapses before treili:-1 piocessing produces bit decision l : jiven initial llr into enhanced llr.> i:i ¦: _ •'.¦¦ -it: a receiver architecture for rsduced-state processing that can be n ;J to im]> ¦:;!.¦- 'I"". DDFSE architecture described above. The state metrics ym(k + l) auu i:.. sets
of -MSI dtoision symbols ck_u {in) that siv Attached to the states are updated by ui;n V'\ As [-.an cl~ the state metrics update comput; .ions, the different candidate metrics x ('i are
i'eri^rai-'^ Tbzic candidate metrics ;.rc used by unit 402 in order to derive the 'riiial LIA.;/" , , ,, L. These initial values are dien combined in unit 403 with the bit u.-c^ons
I
as?rcir;f'-'i with the symbol decision ck_ (m) obtained from a deduction of which ir. • . tate has the best metric. As a result of this combination, the updated LLRs I ,, , t are
generated.
Note that the proposed receiver architecture for the reduced-state approach need ac. re
any memory storage for the initial LLRs or the best candidate melric indices fm (/CH 1) The proposed receiver architecture only requires storage for the state metrics (as is the case for any Viterbi processing) and for the past decision symbolsck_u {pi)- For each state, only the latest Lr decision symbols need to be stored. Hence, the overall memory requirements for the

storage >:' the prsi decision symbols a '¦- "¦'¦' ¦-' be undLf::toou that, by c .n ".: iy selecting the ".ay th- pasi decision sinbols '-';.. v'-\ ¦ ; :-:o:^d ;• ;¦ .,r.ch trellis stak. n ;J possible to implement the update of theso ,s_ .abols using only simple bit-shift and bit-mas!: o,;-iations.
As "/a:; indicated in the introductory portion of this document, the routines described above according to (be present invention for manipulating metrics and soft decisions :-.:. for i;:.;iu'piL, ut iinpiemented in base station 01 a wireless communications netv/ork or d mobile telephone for use in such a network. Figure 5 illustrates a generic structure that can r-oi'. ,-;--nt a mobile telephone, base station or similar receiver in which the invention can be implemented. The receiver 501 comprises an antenna 502 for acquiring wireless signals, an RF section ^03, an analogue-to-digital conversion (ADC) section 504, a dam processor 505 and a memo: "36. In practice, the receiver will contain many other elements but only those necessary "or an explanation of the implementation of the invention have been shown.
Signals received at the antenna 502 are down-converted in frequency and amplified at RF section 503. The signals are then converted into digital signals by ADC section 504 and passed to processor 505. The processor 505 performs the operations necessary to extract and utilise the information payload of the acquired signals, relying on the memory 506 to provide storage of signal values and other data as required. The processor 505 undertakes the routines described above according to the present invention for manipulating metrics and soft decisions, such as those outlined in Figures 3 and 4.
In the preceding discussion, the invention has been described in the context of a demodulation technique implementing equalisation. It will, however, be readily apparent to the sl'illed reader that the invention is equally applicable to the decoding of convolutionaliy encoded signals, in which case the central change is that the symbol memory that defines each tcllis

stale no longer relates to propagation charmo; memory but instead to the memory arrangement utilised for achieving the convolutional ceding of the signal.

•QAIM:
I. App^ai.; for manipulating a reccr <..j sl wliich ooirc>.p.;\ids to a transmit.;. 4 2. Apparatus according to claim 1, wherein the means for producing said initial soft decision
comprises meais for calculating a log likelihood ratio from said candidate metrics to represent
the initial soft decision.
3. Apparatus according to claim 1 or 2, wherein the means for producing said hard decision
comprises means for storing a decision history of metric selections made in calculating metrics through the trellis and means for tracing back through the decision history to produce said hard decision.
4. Apparatus according to claim 1 or 2, wherein the means for processing the trellis to
produce metric? employs a reduced-state technique in which metric selections made in moving
slirrngii [he u.'.llis arc. treated as symbol decisions for calculating metrics further alo: " the uel'-is ; i- . J;; means for producing sen ¦ : i decision comprises means for selecting one of said symbol decisions to provide said hard decision.
5. Appatalns aecoalini] to any p;eceding claim, wherein the means for modifying said initial
.¦¦ i'l ' '¦.:.- ¦ ' r. : ¦ :r:v; icr s.'. mdiiionni on the sttue of said hard der/'j. cue

24
6. Anpnrofar. according to any preeedirr_; ¦. ini, wherein the means for modifying said initial
soft decision comprises means for selecting, conditional on the polarity of said initial soft
decision, one of several values for srM .-.PH ¦>'ed soft decision.
7. A method of manipulating a received signal which corresponds to a transmitted Hgna!
comprisin:; a series of bits, the method c)uprising a step of processing the received signal
using a Vitcrbi algorithm to produce cr;ic'i.;"te metrics for states arranged in a trellis and to
make sek^inns from smongst the candidate metrics to provide metrics for said states, a r-:.;p of
producing an initial soft decision for ?. b:( -f said series from said candidate metrics, a step of
producing u hard decision for said bit from a metric selection corresponding to reception of
. '. ' : ' ¦, ¦:;\.v:>!if/ing the iru

L>.
v m..., i'i accoiduig to claim 7, wbi.uin the step of producing said initial soft decision con:]'.;:. ¦ ..¦ si^p of calculating a log likelihood ratio from said candidate metrics to "cpresent
9. A T. "iiod according to claim 7 or S, wl^rein the step of producing said hard decision
con:, :i;. ¦¦. siv.p of siLiing a decision hisuj-y oi metric selections made in calculating - ¦¦ " .-s
tlirotigh 1'... trc'lis and a step of tracing Kc!: tlirough the decision history to produce s- ¦ ' 'id
uec t IVJJ ¦.
10. A method according to claim 7 or 8, wherein the step of processing the trellis to produce
net/- s • ":oloys a reduced-state technique in which metric selections made in moving thr^'di
the trellis are treated as symbol decisions for calculating metrics further along the treili; id
the step of producing said hard decision comprises a step of selecting one of said symbol
decisions to provide said hard decision.

11. A method according to any one of ckmn 7 to 10, wherein the step of modifying said initial
soft decision comprises a step of selecting, conditional on the state of said hard decision, one
of several values for said enhanced s<:r rr-icn.> 12. A method accoidir.?, .:> a::y one oi ciai::::; 7 to 11, wherein the step of modifying ?aid initial
soft decision comprises a step of selec.inc. conditional on the pciarity of said mind: 'oft
decision, oiio O'L several ( aiues for said tjj^w^^d soft decision.
13. /pp.r:'1 r '¦-r man'rubting a received ritual using a Viterbi algorithm to produce metrics
ff»i stall.'; '.;\ J trellis, wherein said r- uvrd signal corresponds to a transmitted signal
'. "-'-•-¦' ¦!.¦;•' ; ;•.;"¦(,:¦> of 'Si'is and the tpp comprises means for performing an upoa -. of
:• : L.rcopi...'. . -1 by cii:/il:.i.'\'j. ct.kd;i..atc niUiKJ .. . . .:;_
:¦;;•.; i v '. .M:..,j a Luncti... ¦¦ . ;>.'.iralit}r oi' ;iv: candidate metrics that correspond
to ¦:. r' ;¦;¦; .n ihot riMc bit is a logical one, oceans for calculating a function of a plurality of the c,i,:' ' : • • •:ics i;iat correspond to a djci;;lon that said bit is a logical zero and means for L,a!c. .L. ¦. ¦ : •Ii.!ciori .c of the values O/'L!. .VO functions.
1-i. /• r.i.j. ¦ ' i) i" ii \)\uLving a reee.v,.. ' i usiug ?¦ Vi.crbi alijorithm to prodi'.j .i,
for slater ,i; : trellis, 'vherein said rc.;c:\'od signal corresponds to a transmitted s ynai comprising a series of bits and the method comprises a step of performing an update of said metti.^ co. .f the candidate metrics that correspond ;o a Jeoi:i;an uK.t said bit is a logical cne, r stop of calculating a function of a pluralih c" 'hs candidate metrics that correspond to a decision that said bit is a logical zero and a step of calculating the difference of the values of the two functions.
15. A program for causing data processm?. apparatus for carrying out the method of any one of
claim.- 7 t<. and> 16. A telephone or a base station comprising the apparatus of any one of claims 1 to 6 and 13.

-!0
i>¦ \\\c_ is .. :¦¦....r;u ' ¦ ' • •: moduUitiii;.1, •¦. > reived signal, the method being substantially as
]-.-¦.' ¦!.• "/¦ - .' '.-ithiefpr:-:!:'" ' accompanying Figures.
h\L,ni iJ;;.; J 4tii day of February, 1 (PHASKAR KLM^R DEY)

A Vierbi trellis processing technique in which soft decisions and ran; decisions arc derived from a received signal and the soft decisions are enhanced by being modified using the hard decisions. A log likelihood ratio for a bit of the received signal can be derived by grouping candidate metrics associated with the decision that the bit has a first state, grouping candidate metre:; associated with the decision that the bit has a second state, applying respective functions to the groups and calculating the difference of the function values.

Documents:

00560-kolnp-2007-correspondence-1.1.pdf

00560-kolnp-2007-form-1-1.1.pdf

00560-kolnp-2007-form-3-1.1.pdf

00560-kolnp-2007-form-5-1.1.pdf

00560-kolnp-2007-p.a.pdf

00560-kolnp-2007-pct request form.pdf

00560-kolnp-2007-priority document others.pdf

00560-kolnp-2007-priority document.pdf

0560-kolnp-2007-abstract.pdf

0560-kolnp-2007-claims.pdf

0560-kolnp-2007-correspondence others.pdf

0560-kolnp-2007-description (complete).pdf

0560-kolnp-2007-drawings.pdf

0560-kolnp-2007-form1.pdf

0560-kolnp-2007-form2.pdf

0560-kolnp-2007-form3.pdf

0560-kolnp-2007-form5.pdf

0560-kolnp-2007-international publication.pdf

0560-kolnp-2007-international search authority report.pdf

560-KOLNP-2007-(19-11-2013)-ABSTRACT.pdf

560-KOLNP-2007-(19-11-2013)-ANNEXURE TO FORM 3.pdf

560-KOLNP-2007-(19-11-2013)-ASSIGNMENT.pdf

560-KOLNP-2007-(19-11-2013)-CLAIMS.pdf

560-KOLNP-2007-(19-11-2013)-CORRESPONDENCE.pdf

560-KOLNP-2007-(19-11-2013)-DESCRIPTION (COMPLETE).pdf

560-KOLNP-2007-(19-11-2013)-DRAWINGS.pdf

560-KOLNP-2007-(19-11-2013)-FORM-1.pdf

560-KOLNP-2007-(19-11-2013)-FORM-2.pdf

560-KOLNP-2007-(19-11-2013)-FORM-3.pdf

560-KOLNP-2007-(19-11-2013)-FORM-5.pdf

560-KOLNP-2007-(19-11-2013)-OTHERS.pdf

560-KOLNP-2007-(19-11-2013)-PETITION UNDER RULE 8(1).pdf

560-KOLNP-2007-(24-09-2012)-CORRESPONDENCE.pdf

560-KOLNP-2007-(26-11-2012)-CORRESPONDENCE.pdf

560-KOLNP-2007-ASSIGNMENT.pdf

560-kolnp-2007-form 18.pdf

560-KOLNP-2007-FORM 26.pdf

560-KOLNP-2007-FORM 6_1.0.pdf

560-KOLNP-2007-FORM 6_1.1.pdf

560-KOLNP-2007-GRANTED-SPECIFICATION-COMPLETE.pdf

560-KOLNP-2007-REPLY TO EXAMINATION REPORT.pdf

abstract-00560-kolnp-2007.jpg


Patent Number 260041
Indian Patent Application Number 560/KOLNP/2007
PG Journal Number 14/2014
Publication Date 04-Apr-2014
Grant Date 31-Mar-2014
Date of Filing 15-Feb-2007
Name of Patentee TTPCOM LIMITED
Applicant Address MELBOURN SCIENCE PARK, CAMBRIDGE ROAD, MELBOURN, ROYSTON, HERTFORDSHIRE SG8 6HQ
Inventors:
# Inventor's Name Inventor's Address
1 FATEMI-GHOMI, NAVID 6, TENBY ROAD, FIRMLEY, SURREY, GU16 8UT,
2 VALADON, CYRIL 139, JACKMANS PLACE, LETCHWORTH, HERTFORDSHIRE SG6 1RG
PCT International Classification Number H03M 13/41
PCT International Application Number PCT/GB2005/003114
PCT International Filing date 2005-08-05
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 0418263.0 2004-08-16 U.K.